A common question about Prometheus is where to start out? Should you instrument your applications? Or maybe try integrating with your existing monitoring?

The best place to start is usually the node exporter. This exports a plethora of machine-related metrics out of the box with no special configuration, including the standard memory/CPU/disk/network metrics that everyone looks for.

What you'll need:

• A machine (bare metal or VM) running Linux that has internet access
• Ability to use the command line on that machine
• A web browser